Model Aim
- Relatively simple model that models trends in RR-TB prevalence over
time, addressing potential bias in who gets tested as well as varied
roll-out of share of population being tested

Model Specification
- State-level hierarchical generalized additive model (GAM) that
models the prevalence of RR-TB positive cases per quarter among incident
TB cases between 2014-2019.
- Models risk of positivity by patient and municipality
characteristics (currently, HIV status, age category, sex, and health
unit)
- Separate models are run for new TB cases, re-entry cases, and
relapsed cases
result ~ s(state, bs = "re") + s(time) + s(time, by = state, id = 1) + age_cat + hiv_status + sex
Model includes:
Random intercept for each state
A different smooth function for time by state with a shared
smoothing parameter
Each state-level smoothing parameter varies around a grand smooth
function for time to allow for pooling across states
Fixed effects for patient-level characteristics (e.g. age, HIV
status, sex) and municipality-level characteristics (e.g. health unit of
diagnosis/treatment, FHS coverage by enrollment)
New Cases
Adjusted Model
Same scale

Varied scale

Comparing Models
- Adjusted - HIV status, sex, age
- Adjusted+ - HIV status, sex, age, health unit
- Post 2015 adjusted - HIV status, sex, age, health unit
## `geom_line()`: Each group consists of only one observation.
## ℹ Do you need to adjust the group aesthetic?

Re-Entry Cases
Adjusted Model
Same scale

Varied scale
### Post 2015

Relapse Cases
Adjusted Model

Post 2015

Going Forward:
- Individual- and municipality-level covariates to add:
- Add health unit as another covariate
- Exploring covariates for: urbanicity, Bolsa Familia Coverage, FHS
Coverage, Household crowding, Treatment abadonment rate per
municipality
- Come up with inclusion rules for observations (e.g. after certain
date, above a certain percent tested)